halalbooking/primary repository overview

⁠Required libs and tools

• PostgreSQL 13.x (If psql can't connect to the server)⁠
• Ruby. Better to install via rbenv⁠ or rvm⁠. See .ruby-version file. Exact version is not present in the .ruby-version file. For example regardless of your local version 3.1.2 or 3.1.1 in .ruby-version it will be 3.1. To create an alias:
  • rbenv ln -s ~/.rbenv/versions/3.1.2 ~/.rbenv/versions/3.1
  • rvm rvm alias create ruby-3.1.0 3.1.2
• node 14.x
• Yarn⁠
• For images: imagemagick

⁠Only macOS (10.13 or later)

If you are running Puma server in clustered mode, web workers will crash due to an issue with "Ethon" library⁠. To work around the issue, disable fork safety:

export OBJC_DISABLE_INITIALIZE_FORK_SAFETY=YES

This variable must be exported before starting Puma server. A good place for that would be .pumaenv (Puma-dev), ~/.profile or equivalent.

⁠Project setup

make setup

Also, you will need a "development" version of the production DB. For that just run

make fetchdb

It will fetch the dump and restore it automatically as halalbooking DB.

⁠ClojureScript

⁠Dev setup

1. Install leiningen⁠.
2. Run lein fig to start building and hot reloading CLJS code.
3. Open any cljs-powered page in a browser.

If you prefer to work with concatenated file (because of performance reasons), you can run

lein cljsbuild auto front-whitespace (or "extranet-whitespace", or both)

It will still watch for changes. If you just want to build file once, use:

lein cljsbuild once front-whitespace (or "extranet-whitespace", or both)

To clean the output directory run lein clean.

⁠Devcards setup

1. Install leiningen⁠.
2. Run lein fig.
3. Open /dev/cards in a browser.

⁠Clojure

Dev setup:

1. Install Clojure CLI⁠.
2. Run clj -M:dev.
3. Type (start) in REPL.
4. Open http://localhost:8090⁠.

⁠Database connection setup

By default PGUSER and PGDATABASE is used to define the connection. Otherwise current shell user will be used as a DB user (same behavior as with psql) and "halalbooking" as a DB name.

If you want to override database name default, you can either specify PGDATABASE globally via export in your .bash_profile or provide it directly in shell PGDATABASE=hb clj -M:dev.

⁠Rails setup

By default Rails stubs all Clojure-backed requests in development. Make sure you have HANDLE_CLOJURE_REQUESTS=proxy set up in your env.

⁠Linting

See https://github.com/borkdude/clj-kondo/blob/master/doc/editor-integration.md⁠ for live linting setup.

⁠Tests

⁠Rspec

Run all tests: bin/rake spec:all

Front tests only: bin/rspec front/spec

⁠Clojurescript tests

CLJS based:

1. Run yarn (first time only).
2. Run make cljs-test-auto (if you want to run tests once, use make cljs-test-once).

Capybara based:

1. Install brew install chromedriver (first time only). (or upgrade it: brew upgrade chromedriver)
2. Run lein fig.
3. Run a Capybara feature test as usual.

⁠Clojure backend

You will need to manually start Clojure backend for some tests. Note different environment (test) and database (halalbooking_test). A custom server port is used (8091) so that you can run it alongside a development server on the default port (8090).

env HTTP_KIT__PORT=8091 PGDATABASE=halalbooking_test clojure -M:test -m hb.core

Run tests:

env HTTP_KIT__PORT=8091 HANDLE_CLOJURE_REQUESTS=proxy rspec front/spec/features/search_filters_spec.rb

⁠Clojure tests

1. Install Clojure CLI⁠.
2. Make sure your test DB is prepared: bin/rake db:test:prepare.
3. Run make clj-test-once (or make clj-test-auto if you want tests being re-run once you save a file).

⁠Screenshot tests

Prepare Install Selenium - npm install -g selenium-standalone selenium-standalone install

Run

To run tests against staging edit .hermione.conf.js file.

1. Start rails server & lein
2. Start Selenium - selenium-standalone start
3. Run tests - yarn run screenshots

⁠Profiling & local benchmarks

Run RAILS_ENV=profiling bin/rails s. Note, this is effectively a production environment, so you have to restart application after each code change.

⁠CircleCI

See .circleci/README.

⁠SVG sprite

1. To update run yarn run svg-sprite
2. To prevent image from adding to sprite add _ right before .svg (e. g. circle_.svg)

⁠Enable CSS live reload

1. Set ENV variable CSS_LIVE_RELOAD=true or run server with it CSS_LIVE_RELOAD=true rails s
2. Run yarn run watch-css

⁠Watch & rebuild CSS docs

1. Run yarn run watch-css

⁠Supported browsers

See .browserlistrc file.

⁠Usage

Server:

• rails s

The rails server command launches a web server named Puma which comes bundled with Rails.

• puma-dev

Alternative to rails s.

Puma-dev: A fast, zero-config development server for OS X and Linux.⁠

Use the following syntax to create a symbolic link in ~/.puma-dev directory :

$ cd ~/.puma-dev && ln -s /path/to/halalbooking hb

Set "hb.test" as APP_HOST in your "~/.bash_profile":

export APP_HOST=hb.test

Console:

• rails c

The console command lets you interact with your Rails application from the command line. You must always restart the console after code changes.

⁠Background jobs

⁠Development

By default all jobs are stored into "queue_classic_jobs" table (QC⁠ adapter), but not executed. You have to start a QC worker in parallel if you want background jobs, such as sending emails, to be executed:

make jobs

If you don't want the worker to process all previously queued jobs, clear all of them with

bin/rake qc:truncate

⁠Testing

Enqueueing jobs is stubbed and thus it can be asserted⁠.

⁠Staging

We don't run recurring jobs⁠ on staging.

⁠Production

All jobs are stored into "queue_classic_jobs" table and picked up by multiple QC workers (see example⁠).

We run recurring jobs via cron⁠.